SkelSeg: Segmentation and Rigging of Raw-Scanned 3D Volume with User-Specified Skeleton

Although RGB-D camera-based scanning has become popular, a raw-scanned 3D model contains several problems that hinder animation such as fused arms and legs. We propose a system that allows a user to generate a rigged 3D mesh from a raw-scanned 3D volume with simple annotations. The user annotates the skeleton structure on the calibrated images captured at the scanning step, and our system automatically segments the raw-scanned volume into parts, generating a skinned 3D mesh based on the user-specified 3D skeleton. We tested our method with several raw-scanned 3D plush toy models, and successfully generated plausible animations.